|Home | About | Journals | Submit | Contact Us | Français|
Glyoxalase II (GLX2, EC 188.8.131.52., hydroxyacylglutathione hydrolase) is a metalloenzyme involved in crucial detoxification pathways. Different studies have failed in identifying the native metal ion of this enzyme, which is expressed with iron, zinc and/or manganese. Here we report that GloB, the GLX2 from Salmonella typhimurium, is differentially inhibited by glutathione (a reaction product) depending on the bound metal ion, and we provide a structural model for this inhibition mode. This metal-dependent inhibition was shown to occur in metal enriched forms of the enzyme, complementing the spectroscopic data. Based on the high levels of free glutathione in the cell, we suggest that the expression of the different metal forms of GLX2 during Salmonella infection could be exploited as a mechanism to regulate the enzyme activity.
The ubiquitous glyoxalase system is composed of two metalloenzymes (glyoxalase I and glyoxalase II), whose physiological role is to catalyze the conversion of methylglyoxal to the corresponding 2-hydroxycarboxylic acids . Glyoxalase I (GLX1, EC 184.108.40.206., lactoylglutathione methylglyoxal lyase) converts a thiohemiacetal (formed from methylglyoxal and glutathione) to S-d-lactoylglutathione (SLG), while glyoxalase II (GLX2) takes the latter as a substrate and catalyzes its hydrolysis to yield D-lactic acid and glutathione (GSH) [2,3] (Figure 1A).
Although GLX2 is able to hydrolyze many different glutathione thioesters, S-d-lactoylglutathione (SLG) is the preferred substrate for this enzyme from most sources including human, yeast, and plants. The glyoxalase enzymatic system is essential for chemical detoxification, and therefore has been identified as a target for the development of anti-cancer and anti-protozoan drugs [4–6].
GLX2 belongs to the metallo-β-lactamase superfamily, which is characterized by an αβ/βα fold containing a conserved motif able to bind up to two metal ions in their active sites. Most hydrolases from this superfamily use ZnII, a redox inactive metal ion for their roles, such as metallo-β-lactamases [7,8], N-Acyl Homoserine (AHL) Lactonases [9,10] and tRNAseZ , whereas the terminal oxidase Rubredoxin:Oxygen Oxidoreductase (ROO) from D. gigas contains a binuclear iron site [12–14]. GLX2, despite exhibiting the same coordination sphere displayed by other hydrolases from this superfamily (Figure 1B), has been an outstanding exception regarding the metal ion dependence with different metal cofactors [15–20].
In general, the wild type forms of the recombinant GLX2 isoenzymes characterized up to now were isolated with varying amounts of iron, zinc and manganese bound to their active sites [15–20]. Surprisingly, the different metal derivatives show comparable high catalytic efficiencies [15–20]. An exception to this observation is human GLX2, which has been reported to be inactive in its Mn-bound form . Despite the existence of three crystallographic structures (2QED, 1QH5, 1XM8) [18,19,21] which have been refined with different metal ions in its active sites, this puzzling promiscuity has precluded the identification of the native metal ion in GLX2.
Most mechanistic proposals so far have considered exclusively the di-ZnII derivative, based on crystallographic, kinetic and spectroscopic studies, and (more recently) theoretical calculations at the quantum level [17,21,22] (Figure 1B). However, one may argue whether this mechanism could be general to all metal derivatives.
Here we report a spectroscopic and kinetic study on GloB, the glyoxalase II from Salmonella typhimurium,  upon addition of substrate S-d-Lactoylglutathione (SLG) or the product glutathione (GSH). We have interrogated either the enzyme purified from rich medium, containing mixed metallic centers, as well as Fe-, Zn- and Mn- enriched forms . We have found that GloB is differentially inhibited by glutathione depending on the bound metal ion, and we provide a structural model for this inhibition mode.
S-d-Lactoylglutathione (SLG) was purchased from Sigma-Aldrich. All chromatographic steps were performed in an Amersham Biosciences liquid chromatography system operating at 4 °C. Metal standards were purchased from Fisher Scientific and were diluted with distilled water. All other chemicals used in this study were purchased commercially and were of the highest quality available.
GLX2 (GloB) from S. typhimurium 14028s recombinantly produced in E. coli, was used for this study. Protein overexpression and purification were performed using the pET32-gloB vector, as previously described . Fractions with GLX2 activity were pooled and dialyzed against 10 mM 4-morpholinepropanesulfonic acid (MOPS) 0.2 M NaCl at pH 7.2. In order to selectively produce metal-enriched enzyme forms, minimal media M9 was employed in culture growth supplemented with metal ions from stock solutions of Fe(NH4)SO4, MnCl2, or ZnCl2, to reach a final concentration of 100 µM in the culture. The minimal media contained 4 g/L D-(+)-glucose (Sigma), 12.8 g/L Na2HPO4.7H2O, 3 g/L KH2PO4, 0.5 g/L NaCl, 1.0 g/L ammonium sulfate, 10 µM CaCl2 and 1 mM MgSO4. The metal ions were added in the moment of induction, when the growth cultures reached an OD600 of 0.6–0.8. The purification protocol is described in ref. . Thus, we obtained the iron-enriched GloB (FeGloB), manganese-enriched GloB (MnGloB) and zinc-enriched GloB (ZnGloB).
Enzyme concentration was determined by measuring the A280nm and using the extinction coefficient ε280nm = 28,030 M−1 cm−1 . This parameter was calculated from amino acid composition information by applying a modified Edelhoch method .
The metal content of the GloB samples was measured using atomic absorption spectroscopy in a Metrolab 250 AA instrument. Purified enzymes were diluted with 10 mM MOPS pH 7.2, to a concentration of 30 µM and analyzed for zinc, manganese and iron. The metal content of GloB produced in minimal medium was the following: FeGloB: 0.86 ± 0.02 eq iron, 0.05 ± 0.02 eq manganese and 0.04 ± 0.01 eq zinc; MnGloB: 0.04 ± 0.01 eq iron, 0.81 ± 0.06 eq manganese and 0.04 ± 0.01 eq zinc; ZnGloB: 0.08 ± 0.01 eq iron, 1.42 ± 0.01 eq zinc and manganese was not detected. The metal content data presented in this paper represent an average from at least three preparations of each growth condition.
NMR spectra were recorded in a Bruker Avance II 600 spectrometer operating at 600.13 MHz. 1H NMR spectra were recorded under conditions to optimize detection of the fast-relaxing paramagnetic resonances, either using the superWEFT pulse sequence [24,25] or water presaturation. Spectra were acquired over large spectral widths with acquisition times ranging from 16 to 80 ms, and intermediate delays from 2 to 35 ms. 1D experiments with solvent presaturation were used to record isotropically shifted signals closer to the diamagnetic envelope. The samples were loaded into Wilmad 5-mm NMR tubes. The substrate or product were prepared as 100 mM stock solutions in 100 mM MOPS at pH 7.2 (pD 6.8). A final 1:10 enzyme to reactant ratio was used to record the spectra in presence of the adducts.
X-band (9.5 GHz) EPR spectra were measured on a Bruker ELEXSYS E580 system (Bruker BioSpin, Germany) at 4.7 K. The temperature was controlled with an Oxford ITC503 cryogenic system. EPR samples containing a convenient amount of protein or the protein to which the substrate or product were added (up to a 1:10 ratio) were frozen by immersion in liquid nitrogen and then placed in the spectrometer rectangular cavity. All EPR data were corrected by subtracting a baseline corresponding to the EPR signal of the buffer. Other acquisition conditions: modulation amplitude, 1 mT; modulation frequency, 100 kHz; microwave power, 4 mW. A typical EPR sample was 1 mM in 10 mM MOPS pH 7.2 buffer.
Spectra of GloB samples in absence and presence of substrate or product in different ratios, were recorded between 240 and 800 nm at 25° C in a Jasco 550 UV-vis spectrophotometer in quartz cuvette of 1 cm of path length. Differential spectra were obtained by subtracting the spectrum of the enzyme in its unbound form. Typical samples of GloB containing different metal ions were up to 1 mM in protein, while the spectra of Fe-enriched, Mn-enriched and Zn-enriched were recorded on samples typically ca. 30 µM.
Inhibition studies were performed on selectively metal-enriched GloB using glutathione, one of the products of SLG hydrolysis. Steady state kinetic studies of purified GloB variants in presence of varying concentrations of GSH (0.2 to 5 mM) were conducted by measuring the rate of hydrolysis of a fixed concentration of SLG (400 µM) at 240 nm (ε240=3,100 M−1 cm−1). A glutathione stock solution was prepared in 100 mM MOPS, pH 7.2. The enzyme (400 nM) and inhibitor were incubated, and then the substrate was added. IC50 values were determined by fitting the data of initial rates to the following equation: (1−vi/vo)*100 = Imax.[I]/(IC50+[I]). Measurements were performed at least by triplicate in a reaction volume of 1 ml of 10 mM MOPS, 0.2 M NaCl pH 7.2 buffer, at 30° C in a Jasco 550 UV-vis spectrophotometer.
GloB was obtained as previously described in Campos-Bermudez et al.. The typical metal content of enzyme samples expressed in rich medium is 0.2 zinc; 0.6 iron and 0.3 manganese per protein. Selectively metallated variants were produced by expressing the protein in minimal medium adequately supplemented with the required metal ion. The behavior of these proteins upon the addition of substrate or product was analyzed by different spectroscopic techniques.
The EPR spectrum of resting-state GloB recorded at 4.7 K (Figure 2A) disclosed the coexistence of different paramagnetic species. As already shown, the different observed signals can be assigned on the basis of spectral simulations . The main components in the EPR spectra are due to two different iron centers. Resonances at geff ~ 4.3 and 9.1 are characteristic of magnetically isolated high-spin FeIII in a rhombic environment [15,18,19,26–28]. The broad feature at g < 2 corresponds to an antiferromagnetically coupled dinuclear center which involves a high-spin (S=5/2) FeIII ion and a high-spin (S=2) FeII ion [15,18,19,28]. The six-line pattern centered around g = 2.0 is due to MnII bound to the active site (Figure 2A).
S-d-lactoylglutathione (SLG, the enzyme substrate) was added to GloB, and the EPR spectrum was recorded after manual mixing and sample freezing. The spectrum (Figure 2B) revealed changes in the resonances stemming from the iron centers at geff ~ 4.3, 9.1 and < 2, while the signals arising from the MnII site could be accounted for in the simulation by the same set of parameters obtained for resting-state GloB, suggesting that the MnII ions are not perturbed. Addition of glutathione (GSH, a reaction product), gave rise to similar changes on the EPR signals corresponding to the iron centers, while those from the MnII sites also remained unperturbed (Figure 2C).
Analysis of the individual features revealed that in the presence of SLG or GSH, the uncoupled signal (geff ~ 4.3) presented higher zero-field splittings (D = 0.55 cm−1 and E/D = 0.285) compared to resting state GloB (D = 0.33 cm−1 and E/D = 0.195). The increase in D-values can be attributed to a change in the coordination sphere of the iron ion, with more electron-donating atoms; while the higher E/D ratio reveals a larger rhombicity in the FeIII center. Both parameters are consistent with the binding of an additional ligand to the FeIII ion [29,30].
Regarding the signal accounting for the coupled center (g < 2), the addition of either SLG or GSH led to lineshape and g-value differences (Figure 2B–C), revealing that the environment of the binuclear center is significantly perturbed in both cases. Spectra of coupled centers are very sensitive to changes in the coordination environment of the binuclear site which can be reflected in the magnitude of the superexchange coupling (J) between the metal centers, and the zero-field splittings (D) of each ion, among other features [31–34]. Effective g-values for these systems could be easily determined assuming a S=1/2 spin system and yielded, in the presence of ligands, 1.95, 1.87, and 1.71 (gav=1.84). Thus, the addition of exogenous ligands (SLG or GSH) is reflected as a significant increase in gav value for GloB in the presence of ligands (gav=1.84) as compared to pure GloB (gav=1.76), which suggests a stronger coupling regime upon ligand binding.
The EPR spectra of Fe-enriched GloB before and after addition of substrate (SLG) or product (GSH) revealed the same spectral features that for the heterogeneous sample of GloB (not shown). The spectra disclosed mononuclear and binuclear Fe centers, and a weak six-line pattern. Double-integration of the spectra showed that this six-line pattern is responsible for less than 5% of the total signal intensity, indicating that this is the result of a very minor contribution from Mn centers in the sample, which is in agreement with metal analyses. The spectra showed that addition of substrate (SLG) or product (GSH), led to the same spectral alterations that those observed in Figure 2 B–C.
The 1H NMR spectrum of GloB was recorded under conditions tailored to optimize detection of the fast relaxing signals close to the paramagnetic metal center . The spectrum of GloB reveals a set of hyperfine-shifted signals attributed to the His and Asp ligands of the metal site (Figure 3) of a weakly antiferromagnetic coupled FeIII/FeII system. Resonances corresponding to an uncoupled FeIII center cannot be detected, since considerably broader lines are expected . The addition of excess substrate (SLG) to GloB, resulted in a new set of resonances sharper than those from the resting-state enzyme (Figure 3B). Addition of GSH to GloB rendered a spectrum very similar to that resulting upon substrate addition (Figure 3C). Since NMR spectra are recorded at room temperature, in contrast with EPR, and NMR acquisition times are rather long (in relation with the GloB kcat= 170 s−1), the time stability of the monitored species suggests that all studied spectra correspond to an enzyme-product adduct. We must recall that the glutathione moiety is present both in the substrate and product molecules (Figure 1A).
To test the hypothesis that the formed species is an enzyme-product complex, the interaction of GloB with SLG and GSH was followed by UV-vis spectroscopy. We initially attempted these experiments with a sample of GloB obtained from expression in rich media, containing Fe, Zn and Mn. The addition of excess SLG resulted in the appearance of a blue color, corresponding to an absorption band centered at 590 nm, with an estimated extinction coefficient of 300 M−1 cm−1 (Figure 4A). This spectral feature vanished after several hours, confirming the formation of a stable adduct with the enzyme. This feature is in good agreement with the presence of a ligand-to-metal charge transfer (LMCT) band of a Cys ligand to a high spin FeIII in an octahedral coordination, reminiscent of those observed in superoxide reductase and model complexes [36,36–38]. An additional intense band at 320 nm is observed in the differential spectrum (Figure 4A), which can be assigned to a (Cys)S-FeII LMCT . Similar UV-vis specta were obtained for Tflp, a ferredoxin-like protein from Thermoanaerobacter tengcongensis  belonging to the metallo-β-lactamase superfamily, where a disulfide bond is near the di-Fe center in the active site.
When GSH was added to GloB, the same spectral features were observed (Figure 4A). These results further confirm that the spectra recorded after addition of SLG correspond to a stable enzyme-product (E-P) adduct, formed as the result of the interaction of the metal site with the Cys-containing glutathione moiety of the substrate (Figure 1A). These data are consistent with the crystal structure of the human enzyme complexed with GSH (1QH5), in which the sulfur atom of this Cys residue is at 2.86 Å from one of the metal ions . The absorption feature at 590 nm observed for the GloB-GSH complex displays a relatively low extinction coefficient, which is consistent with the relatively long S-FeIII bond reported in the crystal structure. The lack of a spectral feature at 280 nm allows us to discard the presence of a Cys-MnII LMCT, i.e., confirming that this adduct is exclusively formed by the iron form of GLX2 .
When the same experiments were performed in the Fe-enriched form, obtained by expression in minimal medium supplemented with FeII, we observed the same LMCT bands upon addition of either SLG and GSH. These experiments unequivocally confirm that the features observed in the mixed GloB sample correspond to the pure iron form. Instead, when similar titrations were conducted in the Mn-enriched and Zn-enriched forms of GloB, no distinctive features could be detected in the differential spectra even upon addition of a 10-fold excess of the exogenous ligands up to 240 nm (Figure 4B). Minor changes in the spectra can be attributed to alterations in the absorption features of the aromatic residues, as results from the fine structure of the differential spectra. These data allow us to discard the formation of a Cys-MnII CT band, that is expected to appear at 280 nm. This is validated by comparison with the spectrum obtained with the Zn-form of the enzyme, which it is not expected to exhibit a LMCT in this spectral range.
The differential behavior of the different metallated forms towards the formation of enzyme-product complexes led us to analyze product inhibition of each metal form of GloB. We obtained the Mn-enriched, Zn-enriched and Fe-enriched forms of GloB, by expression in minimal medium as already described , and we tested their inhibition by glutathione. Fe- and Zn-GloB were inhibited at submillimolar concentrations of glutathione, while the IC50 value for Mn-GloB is 1.5 mM. These results are in agreement with the spectroscopic data, which show strong product binding to the iron-substituted enzyme, and not to the manganese forms. These data also reveal that the behavior of the Zn form resembles that of Fe-GloB.
Glyoxalase II belongs to the metallo-β-lactamase superfamily, which includes several zinc-dependent hydrolases, such as metallo-β-lactamases, endonuclease tRNase Z, AHL lactonase and phosphorylcholine esterase . GLX2 is an exception in that different metal derivatives show sizably high catalytic performances for most isoforms [15–20]. This is valid for the plant and bacterial isozymes characterized so far, since, despite high sequence homology, human GLX2 is inactive in its MnII form . This fact suggests that metal discrimination could play a role in glyoxalase activity.
The proposed mechanism for the di-Zn enzyme assumes binding of S-d-lactoylglutathione substrate to the enzyme, mainly by recognition of the glutathione moiety by active site residues, excluding a direct interaction with the metal ion (Figure 1B) . The metal ion bound to 3 His residues (M1) is proposed to deliver the attacking nucleophile, at the same time stabilizing the development of negative charge in an oxygen atom during formation of a tetrahedral intermediate. After the nucleophilic attack, the M2 site favors S–C bond cleavage by stabilizing the negative charge in the sulfur atom of the glutathione moiety. After protonation and bond cleavage, a lactate and a glutathione moiety remain bound to the M1 and M2 ions, respectively (Figure 1B). A sequential mechanism for product release has been proposed [17,21,22].
Here we report a metal-selective product inhibition, which allow us to obtain some mechanistic insights. By using complementary spectroscopic techniques, we provide evidence supporting that: (1) addition of either substrate (SLG) or product (GSH) to GloB results in the formation of a stable enzyme-product (GloB-GSH) complex; (2) this adduct is formed by the iron variants of the enzyme, and not by the manganese derivatives; (3) a thiol-Fe bond is formed between the Cys moiety of GSH and the metal site; and (4) the dinuclear iron sites show an enhanced antiferromagnetic coupling in the GloB-GSH adduct.
The finding of a stable thiol-Fe bonding interaction can be readily attributed to product inhibition by glutathione, supporting a sequential, ordered mechanism for product release after hydrolysis [17,21,22]. The reported lower KI for glutathione suggests that it is bound more tightly to the active site and therefore is released after D-lactic acid .
EPR spectroscopy reveals that the coupling between the two iron ions is stronger in the enzyme-product complex. This coupling is expected to occur by involving Asp127 as a bridging ligand (Figure 5). One possible rationale for this observation is to assume an equilibrium between the bridged and non-bridged forms in the free enzyme, which is consistent with the variable geometries observed in different crystal structures [15,18,19,21], that would be shifted towards the bridged one upon formation of the EP adduct (Figure 5A). Sharing the Asp127 ligand between the two metal ions upon thiolate binding may compensate the increase in negative charge in the metal coordination sphere. Another possibility is that the sulfur atom from glutathione forms a strong hydrogen bond to the bridging hydroxide, thus imparting more oxo character in the bridging unit, giving rise to a stronger coupling.
Finally, we could also assume that the thiolate group of glutathione is bound to the metal site bridging the two iron ions (Figure 5B), in a position equivalent to the one occupied by the nucleophile in the resting state enzyme. The GloB-GSH complex displays a S-thiol-FeIII and S-thiol-FeII: CT bands (Figure 4). This observation could be accounted for by assuming a bridging glutathione moiety, confirming again that product release is ordered in a sequential fashion.
The most relevant observation regarding this enzyme product complex concerns its metal selectivity. MnII-thiolate bonds are less covalent and weaker than iron-thiolate bonds, providing a rationale for this differential behavior . This is also consistent with the finding that, in general, the MnII forms of different GLX2’s are the most efficient ones [15,16,19,41].
FeII and MnII are regarded as the most important transition metal ions involved in host-bacterial pathogen interactions . The acquisitions of both metal ions, and particularly MnII, is required for intracellular survival and replication of Salmonella enterica serovar typhimurium in macrophages in vitro and for virulence in vivo [44,45]. Thus, fine regulation of metal ions availability in vivo could determine the pathogen survival inside host cells. Since the levels of free glutathione in vivo are high, this differential inhibition mode might be taken as an indication that MnII is the native metal ion, in line with our results. However, additional experimental work is needed to state that.
The glyoxalase system in Salmonella strains is the main defense mechanism to the accumulation of methylglyoxal generated in the phagolysosome during infection . In contrast to other genes involved in methylglyoxal detoxification, that are induced in the Salmonella-containing vacuole during infection (such as those coding GLX1), the expression levels of gloB are constant through the pathogen infection cycle . Based on the evidence herein presented that shows that inhibition by GSH is metal-dependent, the assembly of a specific metal-enriched form of GLX2 could be exploited as a mechanism to regulate the enzyme activity in vivo.
This work was supported by grants from ANPCyT and HHMI to AJV and from PRONEX/FAPESP/CNPq (Grant no 03/09859-2) and CNPq (307102/2006-8) to AJCF. VCB and JMB are recipients of fellowships from CONICET. AJV is a Staff member from CONICET and an International Research Scholar of the Howard Hughes Medical Institute. The 600 MHz NMR spectrometer was purchased with funds from ANPCyT (PME2003-0026) and CONICET.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.